A hard sell

In an ideal world, every single piece of information you ever needed would be available instantly. But that wouldn't be particularly cost-effective, and such a system - if it could even be constructed - would be a nightmare to back up and maintain.

Managing the transparent flow of information onto appropriate media so that it can be accessed when and where needed is the premise of information life cycle management (ILM), a discipline related to, but sitting above, storage. Why should customers do ILM? Simon Jeggo, information manager at IBM Software Group, says if you don't do some form of information management, storing data becomes very expensive.

"I'll give you an example: a company I know has a production database with seven terabytes in it. They haven't archived data off it, there's no ILM policy and now there's 15 years worth of data on it. As a result it's very inefficient, they can't reorganise it and they're having to throw more fast disk at it because there's no way of introducing different types of storage media to the infrastructure."

Data will always last longer than hardware.

Brian Balfe, African business development director, Commvault

Sunil Mahajan, IM services manager at IBM Software Group, points out that the explosion in data and metadata is being driven by customers.

"There was a time when everyone wanted to put in ERP and that was the 'in' thing. Now there's a lot of competition out there because customers are demanding more information and more access to information. Even a soft drink, these days, has calorie counts, where it was manufactured and so on. All that data has to be stored somewhere and the vendors not only need to store it somewhere, but they want to be able to get some intelligence from it. ILM or information as an agenda is a 'to do' that's on every CIO's list."

Brian Balfe, African business development director for Commvault, says ILM's original emergence was technical: it was a technical way of getting a handle around the information storage problem.

"There have been a whole bunch of different rail tracks that people have gone down trying to achieve it, and we've now come full circle by focusing on why they did it in the first place. It's because there's value in the data that's there, but it's useless if you can't get access to it. So the seven that became 70 that became 700 terabytes sitting on the SAN is there, but the underlying cost is about feeding and watering it and backing it up. People now have a strategy that determines how information is managed from when it's born to when they can get rid of it, but they're forced by economic reasons and capacity to do it in a way that doesn't cost three times more than the company earns in a year. Data will always last longer than hardware, and so it will hop across multiple storage platforms over its lifetime. Having a view of what you've got and how you can retrieve it is driving most of the ILM discussions we're having today."

Dina van Zyl, IM consultant at Emerge Africa, says business benefit will be the ultimate driver of adoption.

"We see a lot of businesses that do have a lot of systems and they've implemented maybe a subset of information management because it doesn't always make sense to do it across the whole organisation. Certainly we've come across sites where they don't implement it at all. Yes, it makes sense to do it, but we need to then demonstrate the business benefit to them of being able to retrieve what they need. That is still a challenge: to get the customers to be able to make sense of what they have."

Barry Gill, CTO of Mimecast, says customers have been looking for some time to converge technologies and that ILM was something that showed potential both to reduce the overall storage footprint and be able to use it across multiple applications.

"There's a lot of attempted convergence that's been occurring around ILM so that the information isn't just exploding all the time. Our customer base is driving us to store information so that any application can access it, and there's rich metadata about which application is doing so. Then they can create an ILM policy for each application around a data subset. Previously, it was managed around storage silos with particular silos having their own policies."

Kobus de Beer, enterprise brand manager at Dell, agrees that in many sites, there are storage and information silos across departments and LOBs where there is little communication or overall policy.

"As storage increased, people threw hardware at it, which didn't really solve the problem. We're now seeing a lot more requests from customers for storage assessments to understand what they've got, how they can use what they have better and whether they have the right software."

Philippe Morin, business unit manager at 3fifteen, says information management should be based on culture.

"You need to look at the organisation and its culture: who is working with information, what they're working on and at what stage. ILM and the governance around the way it's created are important. There are knowledge creators and knowledge sponges who consume it. How do they access that information on a daily basis? Whatever collaboration or business cycle you're working through, what needs to be restored and retained?"

ILM policy is all very well for certain kinds of data, but as Gill points out, unstructured data is a problem.

"When you start looking at data from a risk perspective, corporate governance says it will adhere to certain standards so that it will mitigate risks or comply with certain legislation. So it's a very clear-cut guideline and when you're dealing with structured data, it's relatively easy to apply an ILM policy to that data. When you move into the unstructured environment, it becomes a lot more difficult to apply the correct context to the data. E-mail is the perfect example of an unstructured store that can expose an organisation to huge amounts of risk. The chairman of the board may be communicating with his wife on a personal and a business level, for example: he may e-mail her saying, 'hi honey, I'm going to be late for dinner tonight and by the way, here's a copy of the contract that we need to sign'."

Enterprise in this country cannot keep storing everything on first-tier storage.

Simon Jeggo, information manager, IBM

Do organisations take ILM seriously? According to Manfred Gramlich, storage practice lead at Sun Microsystems, very few do.

"There are three questions that need to be answered," he says. "Who, why and how? ILM, frankly, isn't for everyone. Do you think that your mom-and-pop shop around the corner needs ILM? Of course not. So it's limited to the large financial institutions and enterprises in this country. They need to adopt an ILM strategy. Do they do it today? No, they don't. I've had conversations with the large financial institutions and they store everything on tier-one storage and they do backup - and that's all they do. They may do some form of archiving but they don't have an ILM strategy. For the last six years I've been trying to fight for tiered storage architecture, but people don't care. Companies are happy to pay for a tier-one architecture, keep everything on spinning disk and keep some form of backups on tape - but they don't have an ILM strategy. It's not limited to South Africa: I was in Europe for 16 years and very few - if any - companies there have an ILM strategy either."

Jeggo says the ILM approach has changed for historical reasons, but there's considerable urgency to do at least something useful.

"When the industry started talking about ILM, it was about storage. And a lot of organisations started looking at it. But as they looked, they realised it was a lot more than just storing data: it's about the quality of data, accessing data and the value to the business. It's such a big issue that you can't do it all at once. Our enterprises in this country are almost at the point where they have to do something about it; in fact, they're hitting it now. They cannot keep storing everything on first tier. The quality of the data in this country is shocking so they've got to do something about that as well. Where do you start? Some companies have started different projects that cover at least some of the areas they need in the context of an ILM strategy."

We need to demonstrate the business benefit.

Dina van Zyl, IM consultant, Emerge Africa

And therein lies a completely different problem - organisational cohesion in the face of implementation details. Gill says enterprises are especially prone to pulling in different directions.

"Each project tends to be tackled by a different team in different functional areas of the organisation. You could have one team investigating taking the ERP data off tape and into a cloud, and another team looking at e-mail archiving and putting their data into some kind of monolithic structure for a specific application. There's no cohesive strategy from the CxO level."

It's possible that legislation will help. When the Privacy Act becomes law and the ECT Act gets a few more teeth as a result, companies will be forced to look at how quickly they can get their hands on a piece of information.

But for now, there's precious little urgency. Tiered storage? Information life cycle management? Who needs it? More than one large company in this country runs backups every night onto tape while throwing more first-tier disks at its current storage requirements and thinks it has a strategy. But there will come a time, as Jeggo points out, when this kind of set-up will collapse under its own weight. The proliferation of unstructured data has accelerated this trend quite considerably. The large banks are at least trying to do something. The large telcos don't care - yet. They're making too much money. For now, spinning disk manufacturers are making merry and the business reasons for putting in ILM, or at the very least just a straight-tiered storage architecture, will have to be very convincing.

"If you cannot sit down at CxO level and have a discussion about business reasons for having an ILM strategy, you're wasting your time," concludes Gramlich.